Predicted-Wavefront Backprojection for knowledge-aided SAR image reconstruction

Abstract

This paper develops a SAR image reconstruction technique that incorporates knowledge of the scene to be imaged. This knowledge takes the form of predictions of the electromagnetic field propagation throughout the known portion of the scene, a predicted-wavefront map. This technique predicts wavefronts with INSSITE/RFScene based on an electromagnetic scattering model of the known portion. The novel SAR image reconstruction technique incorporates this predicted wavefront map into a filtered-backprojection framework, hence the name Predicted-Wavefront Backprojection. While this paper uses only results from INSSITE/RFScene, it formulates predicted-wavefront backprojection with sufficient generality that it can incorporate wavefront predictions from other electromagnetic-scattering codes, such as the full-wave codes TEMPUS, FISC, and SAF. A possible application for predicted-wavefront backprojection is Persistent Intelligence, Surveillance, and Reconnaissance (PISR) of an urban setting.